This paper introduces a novel multiagent learning algorithm which achieves convergence, targeted optimality against memory-bounded adversaries, and safety, in arbitrary repeated games. Called CMLeS, its most novel aspect is the manner in which it guarantees (in a PAC sense) targeted optimality against memory-bounded adversaries, via efficient exploration and exploitation. CMLeS is fully implemented and we present empirical results demonstrating its effectiveness.

PDF, PS, HTML

In Proceedings of the Twenty-seventh International Conference on Machine Learning (ICML 2010), June 2010.

@InProceedings{chakraborty:icml10, title={Convergence, Targeted Optimality and Safety in Multiagent Learning}, author={Doran Chakraborty and Peter Stone}, booktitle={Proceedings of the Twenty-seventh International Conference on Machine Learning (ICML 2010)}, month={June}, url="http://www.cs.utexas.edu/users/ai-lab?chakraborty:icml10", year={2010} }